HNHacker News
TopNewBestAskShowJobs

ksolanki

226 karma · joined February 22, 2011

Co-founder and CEO at Eyenuk (eyenuk.com); working hard making computer vision save human vision.

Email solanki -AT- eyenuk.com

submissionscomments
ksolanki··on Ask HN: Who is hiring? (February 2017)
Eyenuk, Inc | Senior Software Engineer and Research Engineer | Woodland Hills, CA | Fulltime | Onsite | http://www.eyenuk.com

At Eyenuk we are developing the next generation of retinal diagnostic solutions that leverage state of the art computer vision using deep learning with novel low-level morphological image analysis algorithms. Software/algorithms we develop makes critical clinical recommendations, and the healthcare professionals rely on the code we write to make these life-altering decisions.

We are hiring for:

1) Algorithms research engineer: The algorithms research engineer will conduct applied research in computer vision and deep neural networks, developing and implementing high-throughput image analysis algorithms, working closely with clinicians and expert doctors.

2) Senior software engineer: The senior software engineer will architect, design, develop, maintain, and deploy software that implements high-throughput (think hundreds of thousands of images an hour) image-based diagnostics. Desktop, web, and mobile.

Our stack: python/flask, Angular, Qt, and C++.

Interview process: Intro over skype -> Phone screen -> Onsite interview -> Offer.

Send email to jobs@eyenuk.com to apply (US-based candidates; H1b or OPT ok).

ksolanki··on Deep learning algorithm diagnoses skin cancer as well as seasoned dermatologists
Indeed, I second that the time is now. There are several imaging modalities/organs that provide diagnostic information about a variety of human functions or diseases. For example, retina is a unique organ that is amenable to imaging of central nervous system, cardio-vascular system, and microvasculature without any incision. This allows us to detect, screen, monitor, and predict risk for diseases such as diabetic retinopathy, macular degeneration, glaucoma, and even cardio-vascular risk, Alzheimer's, and stroke.

If this interests you and you have developed the expertise, there is another startup opportunity to explore -- please email me at solanki@eyenuk.com. We are bunch of machine learning PhDs, developing/publishing and commercializing deep learning algorithms for disease/risk identification from retinal photography. We play with millions of retinal images, and it's a lot of fun!

ksolanki··on Ask HN: Who is hiring? (January 2017)
Eyenuk, Inc. | Research/software engineers| Los Angeles, CA, USA, ONSITE | Full-time | http://www.eyenuk.com

Did you know that retina (eye) is the only organ in human body where the cardiovascular system, the central nervous system, the microvasculature, all can be imaged without any incision (in vivo)? Abnormalities in retinal images can potentially provide useful information about clinical and sub-clinical cerebrovascular, cardiovascular, and metabolic health of a person.

At Eyenuk we are developing the next generation of retinal diagnostic solutions that leverage state of the art computer vision using deep learning with novel low-level morphological image analysis algorithms. Software/algorithms we develop makes critical clinical recommendations, and the healthcare professionals rely on the code we write to make these life-altering decisions. We are hiring for:

1) Algorithms research engineer: The algorithms research engineer will conduct applied research in computer vision and deep neural networks, developing and implementing high-throughput image analysis algorithms, working closely with clinicians and expert doctors.

2) Software engineer (Full Stack): The full-stack engineer will architect, design, develop, maintain, and deploy software that implements high-throughput (think hundreds of thousands of images an hour) image-based diagnostics. Both web and mobile.

Our stack: python/flask, javascript, Swift, and C++.

While we offer usual benefits (health, vision, etc.), a couple things to mention.

* Guys/gals with young families will find our work environment very supportive (many of us have kids too)!

* We work very very hard (and smart) to get things done, however, in 2016 we hardly worked a weekend (and still met the deadlines).

Send email to jobs@eyenuk.com to apply (US-based candidates H1b or OPT ok).

ksolanki··on Ask HN: Who is hiring? (December 2015)
Eyenuk, Inc. | Los Angeles, CA | Onsite | Full-Time | Visa (full time) | Software Engineer www.eyenuk.com

Eyenuk is an exciting medical software startup that has developed a powerful retinal image analysis platform for screening of diseases such as diabetic retinopathy and agre-related macular degeneration.

Anyone with diabetes is vulnerable to blindness that progresses without any pain or discomfort. Despite it's preventable nature, the condition, called diabetic retinopathy, is the leading cause of blindness in working age adults. Eyenuk has developed EyeArt, a SaaS offering, that automates the retinal disease screening process via automated analysis of retinal images and providing a screening referral outcome. EyeArt is a clinically validated on over 33K patients providing sensitivity that exceeds that of expert human graders. EyeArt has CE Marking (regulatory approval in Europe) and is on its way to FDA clearance.

You will join a passionate team at Eyenuk comprised of software and image analysis experts. We are hiring for multiple positions in the following two broad categories.

* Full-stack software engineers (Python, Frontend, UI)

* Algorithms engineer (low-level image processing, deep learning)

Would welcome anyone who wants to join our hands: have a young family? Great! Need a visa? We'll do everything legally possible to get you one.

job at eyenuk.com

ksolanki··on How Google Hires
The real "best interview technique" is not so much black-and-white structures-vs-unstructured. It depends a lot on the type of job you are interviewing for. Structured interviews work really really well for engineering jobs, particularly software engineering. However for marketing, sales, finance, assistant, etc. jobs, the unstructured interview questions work quite well. These are the jobs for which "Tell me the time when..." actually works.
ksolanki··on How to throttle the FCC to dial up modem speeds
Pretty sure Gandhi said that (I have read his biography 3 times!), but should point out that this is nothing much to do with Hinduism.
ksolanki··on How to throttle the FCC to dial up modem speeds
This doesn't seem very productive. As Gandhi once said, "Hate the sin not the sinner". This is not even punishing the sinner, who are not your everyday public workers but likely some lobbying corp.
ksolanki··on US burns through all high-skill visas for 2015 in less than a week
Your point actually makes H1B worse than "indentured servitude" of the past. Many (if not most) H1B folks stick around at one company just waiting for a green card. This kind-of means it is involuntary in some sense. However, unlike indentured servitude, there the "young person" does not automatically get a green card and is not "free to leave" at the end by default. By default, at the end, he/she has to leave the US.
ksolanki··on How Expedia Buys Its Way To The Top Of Google
This and the Rap Genius stories go on to show why algorithmic search (aka Google search) is not going to last another ten years. It probably already does, but more and more Google algorithms will include lots of ifs and buts.

There is definitely some room for fresh approaches to content discovery, knowledge indexing, and finding answers. New startup(s) or Google rediscovering itself?

ksolanki··on From PhD to Data Scientist: Tips for Making the Transition
Calling all data scientists -- why not consider working on medical and/or biological data instead?

There is not necessarily less money in these fields, but a much much greater potential impact!

ksolanki··on Show HN: Hide Secret Messages In Facebook Photos With My Chrome Extension
Great work! Which JPEG stego algorithm do you use (if you don't mind sharing)?

Edit: Found the linked paper. Thanks!

ksolanki··on How Apple and Amazon Security Flaws Led to My Epic Hacking
Most of the "security questions" can be answered by looking at the Facebook profile (of the person or his/her friends -- at least some have the info public). A motivated hacker can possibly crack even bank accounts using the facebook profile. The account/security is indeed in a big mess.
ksolanki··on When Bad Theories Happen to Good Scientists
My bad theory is that Thomas Kuhn's theory of "paradigm shift" itself could be an example of bad theory as presented in the original article. As in, there might not be too tough to find a few obvious counter examples, essentially (proving?) the theory wrong...
ksolanki··on The real reason we’re upset about Sparrow’s acquisition
We (most in this discussion) seem to forget that the advent of $10 or $0.99 app actually expanded the market. As in, more people paid money for software than before, thus paying for more indy/non-indy developers.

So let me say this again. The App store, with its $1-5 apps, is a good thing for the developers as well as the users.

ksolanki··on Advice for Marissa Mayer from an ex-Yahoo
I really like and admire #2 Spend $$ in finding great talent all the way down to the front-line product managers and engineers. This will often mean paying some hot-shot 23-year old coder 5x more than what she would make at Google or Facebook and beating ridiculous counter offers. Do it. That one great engineer will be worth more than the five engineers you have on the payroll today.

However I do not know if there is a easy way (a "litmus paper") that could say she is 5x better. In all seriousness, I'd like to ask how to make such a determination. I think this advise looks good on paper, but really really hard to implement.

ksolanki··on Face.com API shutting down
While you are at it, I'd like to point to some popular existing datasets for object recognition:

The labelme dataset: http://labelme.csail.mit.edu/instructions.html This is close to what you were thinking (and I liked) -- having annotation done by public.

A list: http://www.computervisiononline.com/datasets

PASCAL VOC: http://pascallin.ecs.soton.ac.uk/challenges/VOC/voc2011/

CALTECH 256, and several others. Many papers presented in CVPR 2012 (http://www.cvpr2012.org/) were indeed on very large datasets.

Do note that some of these datasets are fairly large (may or may not be as broad though).

The algorithms you use will depend on the complexity you can handle. OpenCV, IPP, and CCV (which was on hacker news few days back) could provide good options for algorithms with a training dataset you choose/create.

ksolanki··on Face.com API shutting down
I really like the idea of "open source"-like project for training data. Such a concept would be applicable to almost any machine learning based project. Although there are academic datasets for most domains, they have their own limitations.

Often the secret sauce is not the algorithms themselves, but the availability of (massive/broad) training data.

ksolanki··on Face.com API shutting down
OpenCV is not too difficult to setup, but is not as accurate w.r.t. face recognition -- it is not OpenCV's fault (in fact, it is an amazing project, thanks to folks who contribute to it!) -- the accuracy has lot to do with the size and breadth of the training data.
ksolanki··on Steve Jobs on Average vs Best Software Developers
I think there is a lot of talk about hiring an "A" developer, but almost nothing on what makes an A developer an A developer. More likely than not, it a continuum, and that a lot of engineers have the potential to be an A, but don't end up getting there due to one reason or another. It would be great to see some discussion on what environment could one provide that would enable A players to thrive and reach their potential.
ksolanki··on Recently Open-Sourced Django Apps from Eldarion
Is there a post/tutorial/anything that explains the best practices in customizing and/or extending these apps into our own projects, using or not using pinax?

Should we create new apps that subclass parts that we want to extend?

ksolanki··on This photograph is free
If say a well known media organization says, "You should let me use this photo for free, because I will give you credit and that might help you get business later"

Why won't the "well-known media organization" pay what the photo is worth? Almost by definition they can afford to. If they still negotiate in this manner, that is an insult. No?

ksolanki··on Show HN: InaPic: Yet Another Photo App that Isn't So. Smart Photo Albums.
Glad you like it. InaPic indeed is built to help organize thousands of photos and make the process seamless. Hoping to receive more feedback from the community!
ksolanki··on Ask HN: Startup promised me a job, then backed out after the internship
Let me address the visa part (with a disclaimer that I am not a lawyer, but have gone through the visa-related ups and downs myself).

1) You CAN come on tourist visa and try to look for a job -- it will give you about 90 days with a max of 180 days at a stretch. Super strictly speaking, tourists are not supposed to look for jobs, but I know people who have done this.

2) H1B visa allows premium processing so you can get it approved in 14 days, if you or your future employer can pay $1225 premium processing fees. You can get more information from uscis.gov -> forms -> premium processing. Let me say this again: H1B process does not have to take several months.

As others have mentioned before, I would suggest you to contact companies that offered you jobs but you declined.

ksolanki··on Racism in Tech
The probable reason why you see 50-80% Indians in the in the team, but few in management or team lead positions is that, they somehow put up with this bias, that they somehow are happy with what they get. Aren't there a majority of agricultural workers of Latino descent, often working at below minimum wage? Why? They put up with it, and are mostly happy to have the job. The fact is, many Indians, especially those who are "new" in the country, are happy to do what it takes, at the salary and position that they get, even if it is below market rate. Both sides are at fault, no doubt.

Then, of course, there are those who go out and take risk to do extraordinary things, like starting companies. There is no dearth of famous companies founded by Indians, is there?

ksolanki··on Steve's Google Platform rant
This is what I am taking home:

1) There IS no perfect product for everyone.

2) Making something a platform has some chance of addressing the above.

Me trying to make sense of it, but meanwhile let me say it aloud again. There IS no perfect product for everyone. Thanks, Stevey.

ksolanki··on Hiding your data in plain sight: USB hardware hiding
Not really trivial to detect steganography. There is whole body of work that focus on statistically undetectable data hiding [1]. Much of it is in confines of research labs, but it is still feasible.

[1] http://scholar.google.com/scholar?hl=en&q=secure+stegano...

ksolanki··on Hiding your data in plain sight: USB hardware hiding
The best way, if you really want to hide data, is steganography. Embed the information into innocent looking pictures and video. The trick, of course, would be to either download decoding software or upload somewhere when you want access the hidden data.
ksolanki··on Bushido (YC S11): An App Store For The Web That Can Kickstart Your Side Projects
I came here to say exactly this.

I can take Jason Kincaid's word for it and sign up, or Bushido can actually tell me something, anything, about what it is.

Jason's article does indicate they could be on to something interesting, though.

ksolanki··on Google paper comparing performance of C++, Java, Scala, and Go [PDF]
We find that in regards to performance, C++ wins out by a large margin. However, it also required the most extensive tuning efforts, many of which were done at a level of sophistication that would not be available to the average programmer.

I am a fan of C++, so I liked the first part (that C++ wins, which is not really surprising). However it was not clear if the sophistication they talk about is regarding the Google's internal data-structures or the list of optimizations listed in Section VI-D? Many of these optimizations are not big surprises to a C++ programmer (where possible use hash_map, vector instead of list, initialize data structure outside of loop and try to reuse, and so on).

Overall, I do greatly appreciate the effort and their sharing the results.

ksolanki··on We need an AirBNB for Mentorship--not $35k a year wasted on college
I agree. That is close to what I argued here -

http://news.ycombinator.com/item?id=2604734

← PreviousPage 2 of 4Next →